Skip to content
Menu
Finabase Learn

Strategies · Mean reversion

Gap fading: buying the open after a gap down, tested

Fading the opening gap on SPY, QQQ and IWM over two years: the exact rules, a backtest with costs, and how one day in April 2025 accounts for the whole gross gain.

On this page
  1. The idea
  2. The rules
  3. What the backtest shows
  4. Where it fails: the losing periods
  5. Compared with its neighbours
  6. How Finabase does it
  7. Frequently asked questions
  8. Summary
  9. Check your understanding

A gap is the jump between yesterday's close and today's open. News arrives overnight, orders pile up, and the first trade of the day prints well away from the last one. "Fading" a gap means betting against it: after a gap down, buy at the open and expect the price to recover some of the drop before the close.

The idea is popular because it is simple and because everyone has seen a chart where it worked. This lesson turns it into exact rules and tests it on two years of daily bars for the three largest index funds, with costs. The result on SPY is a strategy that broke even after costs, lost more often than it won, and owed its entire gross gain to a single day. It is a good example of how a backtest can look acceptable in a table and fall apart when its trades are read one by one.

The idea

There are two stories about what happens after a gap down, and they point in opposite directions.

The fading story says the open is an overreaction. Overnight news is priced in thin pre-market trading, anxious sellers send orders for the open, and the opening price overshoots. Once regular trading brings in more buyers, the price drifts back toward yesterday's close.

The continuation story says the gap is information. Something changed, the open is only the first reaction, and the selling continues through the day.

Both happen. The existing lesson on gap types and gap fills describes how traders try to tell them apart from context: news, volume, where the gap sits in a trend. A mechanical fade does none of that. It buys every gap of a given size and relies on the first story being true more often, or by more, than the second.

The rules

Checked once a day, at the open:

  1. Compute the gap: today's opening price divided by yesterday's closing price, minus one.
  2. If the gap is −0.5% or lower (the fund opens at least half a percent down), buy at the open.
  3. Sell at the close of the same day, whatever has happened.
  4. On every other day, hold cash. No position is ever held overnight.

A mirror-image rule fades gaps up: sell short at the open after a gap of +0.5% or more and buy back at the close. It is reported below as a variation.

SPY daily, March to May 2025: days that opened 0.5% or more below the prior close $500 $550 $600 Apr May
SPY daily, March to May 2025. A triangle marks each day that opened 0.5% or more below the prior close: the days the rule bought at the open and sold at the close.

A worked example

On March 28, 2025 SPY closed at $555.66. On the next trading day, March 31, it opened at $549.83, a gap of −1.05%. The rule bought at $549.83. During the day the fund traded back above the previous close and finished at $559.39, a rise of 1.74% from the open. After a 0.10% cost on the buy and on the sell, the trade gained 1.5%.

Four days later it went the other way. On April 4 SPY opened at $523.67, a gap of −2.43% after the previous day's fall, and closed at $505.28. The fade lost 3.7% in one session.

What the backtest shows

The test, stated in full so it can be checked:

  • Data. SPY daily open, high, low and close from November 18, 2024 to November 17, 2026: 501 bars, the two years of daily history the Finabase chart data provides.
  • Trading. Buy at the day's official opening price, sell at the day's official close.
  • Costs. 0.10% of the amount on the buy and again on the sell, so 0.20% per round trip.
  • The fill assumption. The official open is set by the exchange's opening auction. To get that price an order must be in the auction, placed before anyone knows what the price will be; but the rule needs to know the gap before deciding to trade. In practice a trader would act on the pre-market indication or buy just after the open, at a price that can differ from the auction price in either direction. On gap days the first minutes are the most volatile of the session, so this assumption matters more here than in most tests.
  • Left out. Interest on cash (the rule is in cash every night and 86% of days) and taxes. Dividends do not apply: the rule never holds overnight, so it never receives one.
SPY daily, 2024-11 to 2026-11: fading gaps down of 0.5% or more against buy and hold, 0.10% per trade $1 $1.2 Growth of $1 Buy and hold Gap-down fade (open to close) 0% −10% −20% Fall from the peak Dec Feb Apr Jun Aug Oct Dec Feb Apr Jun Aug Oct
SPY daily, November 2024 to November 2026: fading gaps down of 0.5% or more (buy the open, sell the close) against buy and hold, 0.10% per trade. Top: growth of $1. Bottom: the fall from the highest value so far.
Measure Gap-down fade Buy and hold
Total return −0.1% 24.5%
Return per year (CAGR) 0.0% 11.7%
Deepest fall from a peak −10.9% −19.0%
Longest time below a peak 404 trading days 88 trading days
Days with a trade 14% 100% invested
Trades 68 1
Trades with a gain 44% n/a

After two years and 68 trades the rule was back where it started. Thirty trades gained 1.1% on average and 38 lost 0.8% on average. It reached its high on April 9, 2025 and never got back to it: the last 404 trading days of the test were spent below that peak.

What costs did

Before costs, the 68 fades averaged a gain of 0.21% from open to close, and the rule returned 14.5% in total. A round trip costs 0.20%. The average edge and the cost are the same size, so nothing was left.

This is the arithmetic of any day trade. The move available between one open and one close is small, and the cost is paid in full on both ends every time.

What one day did

The trade list matters more than the average here.

Date Gap Open to close Note
April 3, 2025 −3.44% −1.54% The fall continued
April 4, 2025 −2.43% −3.51% The worst fade of the test
April 7, 2025 −3.18% 3.11% A recovery from a very low open
April 9, 2025 −0.61% 11.18% The best fade of the test, by far
April 10, 2025 −3.00% −1.43% The fall resumed

On April 9, 2025 SPY opened slightly lower and then rose 11.18% by the close, one of the largest one-day rallies in the fund's history. The rule held it because the open happened to be 0.61% down. Take that one trade away and the other 67 lose 9.9% after costs; their average move before costs drops from 0.21% to 0.05%, and the median fade was 0.03%.

An average driven by one observation is not an edge. It is one lucky day and 67 coin flips that cost 0.20% each.

The two halves of the period

Period Gap-down fade Buy and hold
November 2024 – November 2025 6.8% 12.2%
November 2025 – November 2026 −6.4% 11.0%

The first year holds April 9. The second year, without such a day, lost 6.4%.

Neighbouring settings and other funds

Totals over the same two years, after costs, with the result before costs in parentheses:

Rule SPY QQQ IWM
Fade gaps down of 0.3% or more −6.2% (17.1%), 111 days −23.1% (2.0%), 141 days 4.5% (37.8%), 138 days
Fade gaps down of 0.5% or more −0.1% (14.5%), 68 days −18.9% (0.3%), 106 days 4.3% (28.2%), 103 days
Fade gaps down of 1.0% or more −0.3% (5.5%), 28 days −10.7% (−0.6%), 54 days 21.6% (33.3%), 46 days
Fade gaps up of 0.5% or more (short) −10.0% (5.4%), 79 days −15.4% (8.7%), 125 days −11.0% (14.3%), 125 days

On QQQ the fade had no edge even before costs: the average gap-down day moved 0.01% from open to close. On IWM, the small-company fund, fading large gaps down did make money after costs, 21.6% on 46 trades, one of which gained 9.6%. One fund out of three, on 46 trades, with a single large day inside it, is not evidence of a rule that works. Fading gaps up by selling short lost on all three funds after costs, and a real short would also pay to borrow the shares.

Where it fails: the losing periods

  • Trend days. The fade's worst trades are days when the gap was the start of the move: April 4, 2025 (−3.7% after costs), June 5, 2026, when SPY opened 0.63% down at $752.31 and closed at $737.55 (−2.2%), and April 3, 2025 (−1.7%). On such days there is no bounce to sell into, and the rule holds to the close by design.
  • Most gaps did not fill. In only 21 of the 68 gap-down days did SPY trade back up to the previous close at any point in the session. A strategy premised on gaps closing was right about a third of the time.
  • A long flat stretch. From its April 9, 2025 high to the end of the test the rule lost 10.9% over 19 months, a slow leak of small losses and costs.
  • Costs and fills. At 0.20% per round trip the strategy needs the average fade to be much larger than it was. The opening price assumed here is the best case; a market order sent at the open on a gap day can fill well away from it.
  • Taxes and account rules. Every trade is a day trade with a short-term result. In the US, day trading in a margin account falls under FINRA's intraday margin rules, which replaced the pattern day trader rule on June 4, 2026. A broker may still apply the old rule until October 20, 2027.

Compared with its neighbours

Approach Holding period Trades on SPY, two years Before costs After costs
Gap-down fade One session 68 14.5% −0.1%
Two-day RSI rule A few days 29 33.9% 26.5%
Buy and hold The whole period 1 24.5% 24.5%

The two-day RSI rule also buys weakness, but at the close and for several days, so each trade has more room to move for the same cost. The same two years also show where the fund's return actually came from: SPY's average move from close to the next open was 0.035% and from open to close 0.015%. More than two-thirds of the gain arrived overnight, when a day-trading rule is in cash.

How Finabase does it

Finabase has no gap-fading tool and none of its tools place or suggest trades. The product helps with seeing gaps, not with trading them.

  • Chart. Daily candles show each gap as the space between one bar's close and the next bar's open; the link above opens SPY. Intraday ranges show how a gap day developed after the open.
  • Trending. The trending list shows the stocks moving most, including before the regular session opens, which is where a gap first becomes visible.
  • Backtest. The Backtest tool trades indicator and candlestick signals at the close of the bar that produced them, $10,000 per signal, with no trading cost. It has no open-to-close trade and no gap condition, so it cannot reproduce this test. A strategy whose result is decided by costs would also look much better there than it is.
  • Strategy. The Strategy tool replays if/then allocation rules on daily values over two years. It does not trade within a day.

Frequently asked questions

Do gaps always fill?

No. In this test SPY opened 0.5% or more below the prior close 68 times in two years and traded back to that close during the same session on 21 of them. Many gaps do fill eventually, but "eventually" can be weeks later, after a much lower price.

Why did the strategy break even when it gained 14.5% before costs?

Because it made 68 round trips and each cost 0.20%. The average fade gained 0.21% from open to close before costs, so the cost took essentially all of it.

Is fading gaps up any better?

Not in this data. Shorting the open after a gap up of 0.5% or more and covering at the close lost 10.0% on SPY, 15.4% on QQQ and 11.0% on IWM after costs, before any cost of borrowing shares.

Does the size of the gap matter?

Larger gaps were rarer and no more reliable on SPY: fading gaps of 1.0% or more made 28 trades and returned −0.3% after costs. On IWM the larger gaps did better, on 46 trades, one of which gained 9.6%.

Summary

Fading a gap means buying the open after a drop and selling the close. On SPY over the two years to November 2026 the rule made 68 trades, won 44% of them and returned −0.1% after costs, against 24.5% for holding. Before costs it gained 14.5%, and all of that came from one session, April 9, 2025; the other 67 trades lost money. Two-thirds of the gaps did not close the same day. The test also assumes a perfect fill at the opening price. A rule with an average edge the size of its trading cost, carried by a single outlier, has not shown that it works.

Check your understanding

4 questions on this lesson. Each answer comes with the reason, and nothing is scored.

  1. A fund closed at $100.00 yesterday and opens at $99.30 today. What does the gap-down fade do?

    1. Nothing, because the gap is smaller than 1%
    2. Buys at the open and sells at today’s close
    3. Sells short at the open
    Show the answer

    B. Buys at the open and sells at today’s close The open is 0.7% below the prior close, past the 0.5% threshold. The rule buys the open and always sells at the same day’s close.

  2. The 68 fades gained 0.21% on average before costs, and a round trip costs 0.20%. What follows?

    1. The strategy keeps most of its gain
    2. The cost is about as large as the edge, so almost nothing is left
    3. Costs only matter for strategies that hold overnight
    Show the answer

    B. The cost is about as large as the edge, so almost nothing is left A day trade pays the cost on the buy and on the sell for one session’s move. With an average move the same size as the cost, the result after costs was −0.1%.

  3. Removing one trade, April 9, 2025, changes the result from −0.1% to −9.9%. What does that tell you?

    1. The average was driven by a single outlier, so it is not evidence of a repeatable edge
    2. April is the best month for the strategy
    3. The backtest contains an error
    Show the answer

    A. The average was driven by a single outlier, so it is not evidence of a repeatable edge One open-to-close gain of 11.2% carried all 68 trades. The median fade was 0.03%. When a result depends on one observation, the honest reading is that the other trades had no edge.

  4. Why is the opening price a weak point of this backtest?

    1. Opening prices are not published
    2. The rule must know the gap before trading, but the official open is set in an auction that an order has to join in advance
    3. Funds do not trade at the open
    Show the answer

    B. The rule must know the gap before trading, but the official open is set in an auction that an order has to join in advance The test assumes a fill exactly at the official open. A real trader acts on a pre-market indication or buys just after the open, when prices move fastest, and may get a noticeably different price.

Sources and further reading

  1. Finabase chart data — SPY, QQQ and IWM, daily open, high, low and close, November 2024 to November 2026; rule and results computed by Finabase
  2. SEC — Investor Bulletin: Trading Basics (how market and limit orders are filled)
  3. NYSE — Opening and closing auctions
  4. FINRA — Regulatory Notice 26-10: intraday margin standards replace the day-trading margin rules, effective June 4, 2026, phase-in to October 20, 2027